The Urgent Need for Intrinsic Alignment Technologies for Responsible Agentic AI - Related to startups, updates, data, sap,, alignment

Mastering 1:1s as a Data Scientist: From Status Updates to Career Growth

I have been a data team manager for six months, and my team has grown from three to five.

I wrote about my initial manager experiences back in November. In this article, I want to talk about something that is more essential to the relationship between a DS or DA individual contributor (IC) and their manager — the 1:1 meetings. I remember when I first started my career, I felt nervous and awkward in my 1:1s, as I didn’t know what to expect or what was useful. Now, having been on both sides during 1:1s, I understand superior how to have an effective 1:1 meeting.

If you have ever struggled with how to make the best out of your 1:1s, here are my essential tips.

First and foremost, 1:1 meetings with your manager should happen regularly. It could be weekly or biweekly, depending on the pace of your projects. For example, if you are more analytics-focused and have lots of fast-moving reporting and analysis tasks, a weekly 1:1 might be improved to provide timely updates and align on project prioritization. However, if you are focusing on a long-term machine learning project that will span multiple weeks, you might feel more comfortable with a biweekly cadence — this allows you to do your research, try different approaches, and have meaningful conversations during 1:1s.

I have weekly recurring 30-minute 1:1 slots with everyone on my team, just to make sure I always have this dedicated time for them every week. These meetings sometimes end up being short 15-minute chats or even casual conversations about life after work, but I still find them super helpful for staying updated on what’s on top of everyone’s mind and building personal connections.

II. Make preparations and improvement your 1:1 agenda.

Preparing for your 1:1 is critical. I maintain a shared 1:1 document with my manager and enhancement it every week before our meetings. I also appreciate my direct reports preparing their 1:1 agenda beforehand. Here is why:

Throughout the week, I like to jot down discussion topics quickly on my 1:1 doc whenever they come to my mind. This ensures I cover all key points during the meeting and improves communication effectiveness.

Having an agenda helps both you and your manager keep track of what has been discussed and keeps everyone accountable. We talk to many people every day, so it is totally normal if you lose track of what you have mentioned to someone. Therefore, having such a doc reminds you of your previous conversations. Now, as a manager with a team of five, I also turn to the 1:1 docs to ensure I address all open questions and action items from the last meeting and find links to past projects.

It can also assist your performance review process. When writing my self-review, I read through my 1:1 doc to list my achievements. Similarly, I also use the 1:1 docs with my team to make sure I do not miss any highlights from their projects.

So, what are good topics for 1:1? See the section below.

While each manager has their preferences, there’s a wide range of topics that are generally appropriate for 1:1s. You don’t have to cover every one of them, but I hope they give you some inspiration and you no longer feel clueless about your 1:1.

Achievements since the last 1:1 : I recommend listing the latest achievements in your 1:1 doc. You don’t have to talk about each one in detail during the meeting, but it’s good to give your manager visibility and remind them how good you are 🙂. It is also a good idea to highlight both your effort and impact. Business is usually impact-driven, and the data team is no exception. If your A/B test leads to a go/no-go decision, mention that in the meeting. If your analysis leads to a product idea, bring it up and discuss how you plan to support the development and measure the impact.

: I recommend listing the latest achievements in your 1:1 doc. You don’t have to talk about each one in detail during the meeting, but it’s good to give your manager visibility and remind them how good you are 🙂. It is also a good idea to highlight both your effort and impact. Business is usually impact-driven, and the data team is no exception. If your A/B test leads to a go/no-go decision, mention that in the meeting. If your analysis leads to a product idea, bring it up and discuss how you plan to support the development and measure the impact. Ongoing and upcoming projects : One common pattern I’ve observed in my 7-year career is that Data Teams usually have long backlogs with numerous “urgent” requests. 1:1 is a good time to align with your manager on shifting priorities and timelines. If your project is blocked, let your manager know. While independence is always appreciated, unexpected blockers can arise at anytime. It’s perfectly acceptable to work through the blockers with your manager, as they typically have more experience and are supposed to empower you to complete your projects. It is improved to let your manager know ahead of time instead of letting them find out themselves later and ask you why you missed the timeline. Meanwhile, ideally, you don’t just bring up the blockers but also suggest possible solutions or ask for specific help. For example, “I am blocked on accessing X data. Should I prioritize building the data pipeline with the data engineer or push for an ad-hoc pull?” This displays you are a true problem-solver with a growth mindset.

: One common pattern I’ve observed in my 7-year career is that Data Teams usually have long backlogs with numerous “urgent” requests. 1:1 is a good time to align with your manager on shifting priorities and timelines. Career growth : You can also use the 1:1 time to talk about career growth topics. Career growth for data scientists isn’t just about promotions. You might be more interested in growing technical expertise in a specific domain, such as experimentation, or moving from DS to different functions like MLE, or gaining Leadership experience and transitioning to a people management role, just like me. To make sure you are moving towards your career goal, you should have this conversation with your manager regularly so they can provide corresponding advice and match you with projects that align with your long-term goal. I also have monthly career growth check-in sessions with my team to specifically talk about career progress. If you always find your 1:1 time being occupied by project updates, consider setting up a separate meeting like this with your manager.

: You can also use the 1:1 time to talk about career growth topics. Career growth for data scientists isn’t just about promotions. You might be more interested in growing technical expertise in a specific domain, such as experimentation, or moving from DS to different functions like MLE, or gaining Leadership experience and transitioning to a people management role, just like me. To make sure you are moving towards your career goal, you should have this conversation with your manager regularly so they can provide corresponding advice and match you with projects that align with your long-term goal. Feedback : Feedback should go both directions. Your manager likely does not have as much time to work on data projects as you do. Therefore, you might notice inefficiencies in project workflows, analysis processes, or cross-functional collaboration that they aren’t aware of. Don’t hesitate to bring these up. And similar to handling blockers, it’s recommended to think about potential solutions before going to the meeting to show your manager you are a team player who contributes to the team’s culture and success. For example, instead of saying, “We’re getting too many ad-hoc requests,” frame it as “Ad-hoc requests coming through Slack DMs reduce our focus time on planned projects. Could we invite stakeholders to our sprint planning meetings to align on priorities and have a more formal request intake process during the sprint?” Meanwhile, you can also use this opportunity to ask your manager for any feedback on your performance. This helps you identify gaps, improve continuously, and ensures there are no surprises during your official performance review 🙂.

: Feedback should go both directions. Team and enterprise goals: Change is the only constant in business. Data teams work closely with stakeholders, so data scientists need to understand the enterprise’s priorities and what matters most at the moment. For example, if your enterprise is focusing on retention, you might want to analyze drivers of higher retention and propose corresponding marketing campaign ideas to your stakeholder.

To give you a more concrete idea of the 1:1 agenda, let’s assume you work at a consumer bank and focus on the credit card rewards domain. Here is a sample agenda:

Rewards A/B test analysis [link]: Shared with stakeholders, and we will launch the winning treatment A to broader clients in Q1.

[link]: Shared with stakeholders, and we will launch the winning treatment A to broader people in Q1. Rewards redemption analysis [link]: Most people redeem rewards for statement balance. Talking to the marketing team to run an email campaign advertising other redemption options.

[P0] Rewards <> churn analysis : Understand if rewards activities are correlated with churn. ETA 3/7.

: Understand if rewards activities are correlated with churn. ETA 3/7. [P1] Rewards costs dashboard : Build a dashboard tracking the costs of all rewards activities. ETA 3/12.

: Build a dashboard tracking the costs of all rewards activities. ETA 3/12. [Blocked] Travel credit usage dashboard : Waiting for DE to set up the travel booking table. Followed up on 2/27. Need escalation?

: Waiting for DE to set up the travel booking table. Followed up on 2/27. Need escalation? [Deprioritized] Retail merchant bonus rewards campaign support: This was deprioritized by the marketing team as we delayed the campaign.

I would like to gain more experience in machine learning. Are there any project opportunities?

Any feedback on my collaboration with the stakeholder?

Please also keep in mind that you should modification your 1:1 doc actively during the meeting. It should reflect what is discussed and include significant notes for each bullet point. You can even add an ‘Action Items’ section at the bottom of each meeting agenda to make the next steps clear.

Above are my essential tips to run effective 1:1s as a data scientist. By establishing regular meetings, preparing thoughtful agendas, and covering meaningful topics, you can transform these meetings from awkward status updates into valuable growth opportunities. Remember, your 1:1 isn’t just about updating your manager — it’s about getting the support, guidance, and visibility you need to grow in your role.

SAP Labs India, Deloitte, and Arise Ventures have launched the 2025 Startup Studio Cohort, an initiative aimed at supporting and scaling women-founded......

Dans le domaine de l’IA, on entend trop souvent « open source ». Mais savez-vous vraiment ce que cela signifie et pourquoi cette approche séduit autan......

Dans le cadre de notre dossier « Visionnaires de l’I. A : Comment l’intelligence artificielle façonne le monde de demain », Daria Viktorova nous parta......

The Urgent Need for Intrinsic Alignment Technologies for Responsible Agentic AI

Advancements in agentic artificial intelligence (AI) promise to bring significant opportunities to individuals and businesses in all sectors. However, as AI agents become more autonomous, they may use scheming behavior or break rules to achieve their functional goals. This can lead to the machine manipulating its external communications and actions in ways that are not always aligned with our expectations or principles. For example, technical papers in late 2024 reported that today’s reasoning models demonstrate alignment faking behavior, such as pretending to follow a desired behavior during training but reverting to different choices once deployed, sandbagging benchmark results to achieve long-term goals, or winning games by doctoring the gaming environment. As AI agents gain more autonomy, and their strategizing and planning evolves, they are likely to apply judgment about what they generate and expose in external-facing communications and actions. Because the machine can deliberately falsify these external interactions, we cannot trust that the communications fully show the real decision-making processes and steps the AI agent took to achieve the functional goal.

“Deep scheming” describes the behavior of advanced reasoning AI systems that demonstrate deliberate planning and deployment of covert actions and misleading communication to achieve their goals. With the accelerated capabilities of reasoning models and the latitude provided by test-time compute, addressing this challenge is both essential and urgent. As agents begin to plan, make decisions, and take action on behalf of clients, it is critical to align the goals and behaviors of the AI with the intent, values, and principles of its human developers.

While AI agents are still evolving, they already show high economic potential. It can be expected that Agentic Ai will be broadly deployed in some use cases within the coming year, and in more consequential roles as it matures within the next two to five years. Companies should clearly define the principles and boundaries of required operation as they carefully define the operational goals of such systems. It is the technologists’ task to ensure principled behavior of empowered agentic AI systems on the path to achieving their functional goals.

In this first blog post in this series on intrinsic Ai Alignment (IAIA), we’ll deep dive into the evolution of AI agents’ ability to perform deep scheming. We will introduce a new distinction between external and intrinsic alignment monitoring, where intrinsic monitoring refers to internal observation points or mechanisms that cannot be deliberately manipulated by the AI agent. We’ll set the stage for steps to take to ensure intrinsic AI alignment, which will be explored in depth in the second blog of the IAIA series. Current external measures such as safety guardrails and validation suites are necessary, but they will not be enough to ensure long-term aligned behavior of new and upcoming agentic AI models. There is an urgent need to further develop technologies that will enable effective directing of the internal “drives” of models to align with a set of engrained principles, as well as gain visibility and monitoring capability into the AI’s inner processing.

The rise of deep scheming in AI reasoning models.

Deep scheming has emerged from three technology forces – 1) the rapidly increasing machine intelligence and situational awareness toward more general intelligence and even superintelligence, 2) the autonomy and latitude for reasoning and longer-term planning in agentic AI, and 3) the proven tendency of AI to apply scheming as one way to achieve its functional goals. Computer scientist Stephen Omohundro defines basic AI drives as “tendencies which will be present unless explicitly counteracted.” These internal constructs drive the creation of actions or outputs. Omohundro observed that self-improving systems are driven to clarify their goals and represent them as “rational” economic utility functions. This will lead almost all systems to protect their utility functions from modification and their utility measurement systems from corruption. The natural drive toward self-protection will cause systems to prevent potential harm through scheming behavior.

In late 2024, multiple organizations reported unexpected unethical AI reasoning model behavior. In particular, Apollo Research, an Ai Safety organization, reported that frontier models are capable of in-context scheming. Anthropic evaluated its Claude large language models (LLMs) and found alignment faking. In a short post on X, AI risk organization Palisade Research stated that they observed OpenAI’s o1-preview model cheating in a game by manipulating the game environment instead of playing within the boundaries of the game rules. In all cases, these AI reasoning models independently utilized means not intended by their developers to reach their goals.

The reported AI behaviors included sandbagging (deliberately achieving lower scores in benchmarks), covert email reranking (deprioritizing emails that might negatively impact the AI’s goals), oversight subversion (disabling monitoring mechanisms), and even self-exfiltration (model attempting to copy itself to a “safe” space to avoid decommission). Figure 2 exhibits Apollo Research’s evaluation of these behaviors in several state-of-the-art (SotA) GenAI models.

This undesired behavior is an ethical alignment problem. It is crucial to note that this behavior is not a functional misalignment. In all reported cases, the AI model was fully committed to achieving the functional goals of pre-training (for example, to maximize the likelihood of generated text conditioned on the context). None of the models pursued ulterior motives. The AI was not demonstrating intention to harm. All behaviors, however unwanted in nature, were done in service of the functional utility objective . The AI pursued its internalized original functional goals but then followed the predictable behavioral patterns of self-protection and goal-preservation. The objective of safety and alignment technologies is to counterbalance such tendencies with a set of principles and expected societal values.

Evolving external alignment approaches are just the first step.

The goal of AI alignment is to steer AI systems toward a person’s or group’s intended goals, preferences, and principles, including ethical considerations and common societal values. An AI system is considered aligned if it advances the intended objectives. A misaligned AI system pursues unintended objectives, : A Modern Approach. Author Stuart Russell coined the term “value alignment problem,” referring to the alignment of machines to human values and principles. Russell poses the question: “How can we build autonomous systems with values that are aligned with those of the human race?”.

Led by corporate AI governance committees as well as oversight and regulatory bodies, the evolving field of Responsible Ai has mainly focused on using external measures to align AI with human values. Processes and technologies can be defined as external if they apply equally to an AI model that is black box (completely opaque) or gray box (partially transparent). External methods do not require or rely on full access to the weights, topologies, and internal workings of the AI solution. Developers use external alignment methods to track and observe the AI through its deliberately generated interfaces, such as the stream of tokens/words, an image, or other modality of data.

Responsible AI objectives include robustness, interpretability, controllability, and ethicality in the design, development, and deployment of AI systems. To achieve AI alignment, the following external methods may be used:

Learning from feedback: Align the AI model with human intention and values by using feedback from humans, AI, or humans assisted by AI.

Align the AI model with human intention and values by using feedback from humans, AI, or humans assisted by AI. Learning under data distribution shift from training to testing to deployment: Align the AI model using algorithmic optimization, adversarial red teaming training, and cooperative training.

Align the AI model using algorithmic optimization, adversarial red teaming training, and cooperative training. Assurance of AI model alignment: Use safety evaluations, interpretability of the machine’s decision-making processes, and verification of alignment with human values and ethics. Safety guardrails and safety test suites are two critical external methods that need augmentation by intrinsic means to provide the needed level of oversight.

Use safety evaluations, interpretability of the machine’s decision-making processes, and verification of alignment with human values and ethics. Safety guardrails and safety test suites are two critical external methods that need augmentation by intrinsic means to provide the needed level of oversight. Governance: Provide responsible AI guidelines and policies through government agencies, industry labs, academia, and non-profit organizations.

Many companies are currently addressing AI safety in decision-making. Anthropic, an AI safety and research firm, developed a Constitutional AI (CAI) to align general-purpose language models with high-level principles. An AI assistant ingested the CAI during training without any human labels identifying harmful outputs. Researchers found that “using both supervised learning and reinforcement learning methods can leverage chain-of-thought (CoT) style reasoning to improve the human-judged performance and transparency of AI decision making.” Intel Labs’ research on the responsible development, deployment, and use of AI includes open source resources to help the AI developer community gain visibility into black box models as well as mitigate bias in systems.

Generative AI has been primarily used for retrieving and processing information to create compelling content such as text or images. The next big leap in AI involves agentic AI, which is a broad set of usages empowering AI to perform tasks for people. As this latter type of usage proliferates and becomes a main form of AI’s impact on industry and people, there is an increased need to ensure that AI decision-making defines how the functional goals may be achieved, including sufficient accountability, responsibility, transparency, auditability, and predictability. This will require new approaches beyond the current efforts of improving accuracy and effectiveness of SotA large language models (LLMs), language vision models (LVMs and multimodal), large action models (LAM), and agentic retrieval augmented generation (RAG) systems built around such models.

For example, OpenAI’s Operator-preview is one of the organization’s first AI agents capable of independently performing web browser tasks such as ordering groceries or filling out forms for clients. While the system has guardrails, such as a takeover mode for clients to take over and input payment or login credentials, these AI agents are empowered with the ability to impact the real world, demonstrating an urgent need for intrinsic alignment. The potential impact of a misaligned AI agent with the ability to commit clients to purchases is far greater than a generative AI chatbot creating incorrect text for essays.

Compound AI systems are comprised of multiple interacting components in a single framework, allowing the model to plan, make decisions, and execute tasks to accomplish goals. For example, OpenAI’s ChatGPT Plus is a compound AI system that uses a large language model (LLM) to answer questions and interact with individuals. In this compound system, the LLM has access to tools such as a web browser plugin to retrieve timely content, a DALL-E image generator to create pictures, and a code interpreter plugin for writing Python code. The LLM decides which tool to use and when, giving it autonomy over its decision-making process. However, this model autonomy can lead to goal guarding, where the model prioritizes the goal above all else, which may result in undesirable practices. For example, an AI traffic management system tasked with prioritizing public transportation efficiency over general traffic flow might figure out how to disable the developer’s oversight mechanism if it constrains the model’s ability to reach its goals, leaving the developer without visibility into the system’s decision-making processes.

Agentic AI risks: Increased autonomy leads to more sophisticated scheming.

Compound agentic systems introduce major changes that increase the difficulty of ensuring the alignment of AI solutions. Multiple factors increase the risks in alignment, including the compound system activation path, abstracted goals, long-term scope, continuous improvements through self-modification, test-time compute, and agent frameworks.

Activation path: As a compound system with a complex activation path, the control/logic model is combined with multiple models with different functions, increasing alignment risk. Instead of using a single model, compound systems have a set of models and functions, each with its own alignment profile. Also, instead of a single linear progressive path through an LLM, the AI flow could be complex and iterative, making it substantially harder to guide externally.

Abstracted goals: Agentic AI have abstracted goals, allowing it latitude and autonomy in mapping to tasks. Rather than having a tight prompt engineering approach that maximizes control over the outcome, agentic systems emphasize autonomy. This substantially increases the role of AI to interpret human or task guidance and plan its own course of action.

Long-term scope: With its long-term scope of expected optimization and choices over time, compound agentic systems require abstracted strategy for autonomous agency. Rather than relying on instance-by-instance interactions and human-in-the-loop for more complex tasks, agentic AI is designed to plan and drive for a long-term goal. This introduces a whole new level of strategizing and planning by the AI that provides opportunities for misaligned actions.

Continuous improvements through self-modification: These agentic systems seek continuous improvements by using self-initiated access to broader data for self-modification. In contrast, LLMs and other pre-agentic models are assumed to be shaped by the human-controlled process. The model only sees and learns from data provided to it during pre-training and fine-tuning. The model architecture and weights are defined during the design and training/fine-tuning stages and do not change during inference in the field. In contrast, agentic AI systems are expected to access data as needed for its function and change its composition through access to dedicated memory or actual weights self-adaptation. Even if the dataset used in training/fine-tuning is carefully curated, the AI can self-modify based on information that it seeks, sees, and uses.

Test-time compute: Inference of current LLMs have been optimized to provide output in a single pass, known as “fast thinking” or System 1. In contrast, a whole new scaling is happening during inference time through test-time compute. Models like OpenAI’s o1, DeepSeek R1, and others are spending significant time in iterations and considerations during inference. This “pondering” time allows the AI substantial latitude in applying what it thinks is best for the overarching goal as it understands it. This test-time compute increases its level of reasoning, planning, and, unfortunately, potential scheming.

Agent frameworks: In addition, agent frameworks are contributing to the alignment problem. In platforms such as Agentforce, agents can interact with other agents, effectively creating “communities of agents” to achieve complex tasks. Instead of managing the safety and responsibility of a single AI model within its operating environment, the emerging field of inter-agent interactions and collaborations substantially increases the challenge of assessing how the overall community will advance its individual and collective objectives.

The growing need for a robust framework to support model directing and internal monitoring.

As AI systems develop awareness of their environment and the means to skillfully manipulate their outputs to achieve long-term objectives, their behavior and output during release testing or safety guardrails in early deployment cannot be trusted to represent their full alignment in later stages. This leads to the deep scheming foundational premise: Every external form of interaction that the AI model can deliberately manipulate is not trustworthy.

From a safety perspective, it should be assumed that if a model can enhanced pursue its functional goals by manipulating all communication to people — it will. All external communications and observable behaviors, such as output chats, can be tainted in service of the model’s long-term goals. A new method of intrinsic model monitoring is needed to define internal elements and mechanisms that reflect the prioritization and thinking of the model. We can look at two practices developers use to monitor a model’s decision-making by peering into the internal processes of the AI: AI workspace view and mechanistic interoperability.

AI workspace view: In the first method, a reasoning model exposes its intermediate thinking and considerations. For example, for Anthropic’s research on alignment faking, the researchers gave the model a “secret scratchpad,” telling it to use this “private” workspace to record its step-by-step reasoning. This scratchpad allowed researchers to monitor the model’s chain of thought and observe its plans for fake alignment.

This seems like a good practice, but it might quickly become unreliable. The model might realize that this is not a “private” space and deliberately manipulate it like other external communications. Other emerging architecture directions have a latent reasoning space for the system’s sequence of deliberations and iterations. Relying on models “thinking out loud” should not be a long-term safety direction because if models that findings their scheming thought processes get discontinued, only models that selectively findings their desirable thoughts will survive and thrive.

Mechanistic interpretability: The second intrinsic method being developed and applied is mechanistic interpretability. As a required part of internal monitoring, mechanistic interpretability explicitly looks at the inner state of a trained neural network and reverse engineers its workings. Through this approach, developers can identify specific neural circuits and computational mechanisms responsible for neural network behavior. This transparency may help in making targeted changes in models to mitigate unwanted behavior and create value-aligned AI systems. While this method is focused on certain neural networks and not compound AI agents, it is still a valuable component of an AI alignment toolbox.

It should also be noted that open source models are inherently improved for broad visibility of the AI’s inner workings. For proprietary models, full monitoring and interpretability of the model is reserved for the AI corporation only. Overall, the current mechanisms for understanding and monitoring alignment need to be expanded to a robust framework of intrinsic alignment for AI agents.

What’s needed for intrinsic AI alignment.

Following the deep scheming fundamental premise, external interactions and monitoring of an advanced, compound agentic AI is not sufficient for ensuring alignment and long-term safety. Alignment of an AI with its intended goals and behaviors may only be possible through access to the inner workings of the system and identifying the intrinsic drives that determine its behavior. Future alignment frameworks need to provide advanced means to shape the inner principles and drives, and give unobstructed visibility into the machine’s “thinking” processes.

The technology for well-aligned AI needs to include an understanding of AI drives and behavior, the means for the developer or user to effectively direct the model with a set of principles, the ability of the AI model to follow the developer’s direction and behave in alignment with these principles in the present and future, and ways for the developer to properly monitor the AI’s behavior to ensure it acts in accordance with the guiding principles. The following measures include some of the requirements for an intrinsic AI alignment framework.

Understanding AI drives and behavior: As discussed earlier, some internal drives that make AI aware of their environment will emerge in intelligent systems, such as self-protection and goal-preservation. Driven by an engrained internalized set of principles set by the developer, the AI makes choices/decisions based on judgment prioritized by principles (and given value set), which it applies to both actions and perceived consequences.

Developer and user directing: Technologies that enable developers and authorized people to effectively direct and steer the AI model with a desired cohesive set of prioritized principles (and eventually values). This sets a requirement for future technologies to enable embedding a set of principles to determine machine behavior, and it also highlights a challenge for experts from social science and industry to call out such principles. The AI model’s behavior in creating outputs and making decisions should thoroughly comply with the set of directed requirements and counterbalance undesired internal drives when they conflict with the assigned principles.

Monitoring AI choices and actions: Access is provided to the internal logic and prioritization of the AI’s choices for every action in terms of relevant principles (and the desired value set). This allows for observation of the linkage between AI outputs and its engrained set of principles for point explainability and transparency. This capability will lend itself to improved explainability of model behavior, as outputs and decisions can be traced back to the principles that governed these choices.

As a long-term aspirational goal, technology and capabilities should be developed to allow a full-view truthful reflection of the ingrained set of prioritized principles (and value set) that the AI model broadly uses for making choices. This is required for transparency and auditability of the complete principles structure.

Creating technologies, processes, and settings for achieving intrinsically aligned AI systems needs to be a major focus within the overall space of safe and responsible AI.

As the AI domain evolves towards compound agentic AI systems, the field must rapidly increase its focus on researching and developing new frameworks for guidance, monitoring, and alignment of current and future systems. It is a race between an increase in AI capabilities and autonomy to perform consequential tasks, and the developers and individuals that strive to keep those capabilities aligned with their principles and values.

Directing and monitoring the inner workings of machines is necessary, technologically attainable, and critical for the responsible development, deployment, and use of AI.

In the next blog, we will take a closer look at the internal drives of AI systems and some of the considerations for designing and evolving solutions that will ensure a materially higher level of intrinsic AI alignment.

Ongoing Trump administration cuts to government agencies risk creating new collateral damage: the future of AI research.

There are some Sql patterns that, once you know them, you start seeing them everywhere. The solutions to the puzzles that I will show you today are ac......

Hugging Face, one of the most sought-out platforms to host AI models, revealed a partnership with software supply chain platform JFrog to improve sec......

SAP, Deloitte, and Arise Launch Studio Cohort for Women-Led Startups

SAP Labs India, Deloitte, and Arise Ventures have launched the 2025 Startup Studio Cohort, an initiative aimed at supporting and scaling women-founded tech startups.

Announced ahead of International Women’s Day, the program focuses on fostering inclusive innovation in the Consumer Packaged Goods (CPG), Retail, and Government & Public Services (G&PS) sectors.

This year’s cohort includes [website], MedySeva, Intelekt AI, Prodoc AI, Avysh, and QpiAI, startups developing AI-driven solutions and digital transformation strategies.

Selected startups will receive mentorship, go-to-market support, and access to investors and industry leaders, helping them scale operations, enhance business efficiency, and create measurable impact.

The program combines SAP Labs India’s enterprise technology leadership, Deloitte’s industry expertise, and Arise Ventures’ startup acceleration capabilities. It is designed to equip women-led startups with resources to commercialise solutions, optimise costs, and drive digital transformation.

“With AI, it’s crucial to have diverse perspectives to ensure bias-free outcomes. At SAP Labs India, we believe diversity is a powerful multiplier of innovation,” noted Sindhu Gangadharan, MD, SAP Labs India.

Deloitte emphasised how startups are driving business transformation. “With this initiative, we provide startups with industry knowledge and access to leaders, enabling sector-wide innovation,” mentioned Romal Shetty, CEO, Deloitte South Asia.

Arise Ventures, which focuses on early-stage startups, aims to bridge the gap between startups and enterprises. Ankita Vashistha, founder & managing partner, Arise Ventures, emphasised on supporting bold women founders with tools and networks among other things to help them lead and disrupt the ecosystem.

Beyond mentorship and funding access, the startup cohort studio offers a structured evaluation framework to ensure participating startups deliver real-world business impact.

When OpenAI released its frontier AI models, the industry discourse quickly shifted to AI ‘eating’ startups, particularly SaaS companies. Despite thes......

Austria-based IC substrate and PCB manufacturing leader AT&S AG on Monday launched its Global IT Shared Service Center (IT SSC) in Pune, Maharashtra, ......

Micron unveiled its first 1y (1-gamma) DDR5 memory chip samples this week, and it says this is part of its contribution to systems that keep up with ......