Decisions such as which movie to watch next, which song to listen to, or which product to buy online, are increasingly influenced by recommender systems and user models that incorporate information on users' past behaviours, preferences, and digitally created content. Machine learning models that enable recommendations and that are trained on user data may unintentionally leverage information on human characteristics that are considered vulnerabilities, such as depression, young age, or gambling addiction. The use of algorithmic decisions based on latent vulnerable state representations could be considered manipulative and could have a deteriorating impact on the condition of vulnerable individuals. In this paper, we are concerned with the problem of machine learning models inadvertently modelling vulnerabilities, and want to raise awareness for this issue to be considered in legislation and AI ethics. Hence, we define and describe common vulnerabilities, and illustrate cases where they are likely to play a role in algorithmic decision-making. We propose a set of requirements for methods to detect the potential for vulnerability modelling, detect whether vulnerable groups are treated differently by a model, and detect whether a model has created an internal representation of vulnerability. We conclude that explainable artificial intelligence methods may be necessary for detecting vulnerability exploitation by machine learning-based recommendation systems.

FIGURE 1: LMU seal (left) style-transferred to Van Gogh's Sunflower painting (center) and blended with the prompt - Van Gogh, sunflowers - via CLIP+VGAN (right). In the last few years, there have been several breakthroughs in the methodologies used in Natural Language Processing (NLP) as well as Computer Vision (CV). Beyond these improvements on single-modality models, large-scale multimodal approaches have become a very active area of research. In this seminar, we reviewed these approaches and attempted to create a solid overview of the field, starting with the current state-of-the-art approaches in the two subfields of Deep Learning individually. Further, modeling frameworks are discussed where one modality is transformed into the other Chapter 3.1 and Chapter 3.2), as well as models in which one modality is utilized to enhance representation learning for the other (Chapter 3.3 and Chapter 3.4). To conclude the second part, architectures with a focus on handling both modalities simultaneously are introduced (Chapter 3.5). Finally, we also cover other modalities (Chapter 4.1 and Chapter 4.2) as well as general-purpose multi-modal models (Chapter 4.3), which are able to handle different tasks on different modalities within one unified architecture.

The continuing development of AI systems represents a profound achievement of the digital age that brings with it tremendous opportunities. In fact, many in the creative industry are already using or plan to use AI for the creation of a wide range of works that benefit society. But as with many advances in technology, these new opportunities come with challenges. This licensing activity is evidence of existing markets for TDM. It is important that the conditions of those licenses are respected and that they are not undermined by new exceptions that excuse unauthorized uses.

The short answer is no. However, there are numerous aspects of strategists' work where AI and advanced analytics tools can already bring enormous value. Yuval Atsmon is a senior partner who leads the new McKinsey Center for Strategy Innovation, which studies ways new technologies can augment the timeless principles of strategy. In this episode of the Inside the Strategy Room podcast, he explains how artificial intelligence is already transforming strategy and what's on the horizon. This is an edited transcript of the discussion. For more conversations on the strategy issues that matter, follow the series on your preferred podcast platform.

Following multiple amendments and discussions, the EU Member States – the Council of the EU – approved a compromise version of the proposed Artificial Intelligence Regulation (AI Act) on December 6, 2022. Once adopted, the AI Act will be the first horizontal legislation in the EU to regulate AI systems, introducing rules for the safe and trustworthy placing on the EU market of products with an AI component. The Regulation's extraterritorial scope (i.e., application to providers and users outside the EU when the output produced by the system is used in the EU) and its exceptionally high fines of the higher of up to €30 million or up to 6 % of the company's total worldwide annual turnover for the preceding financial year, are expected to shape the regulatory requirements outside of the EU borders as has been the case with the European General Data Protection Regulation (GDPR). The first proposal for an AI Act was published by the European Commission (Commission) in April 2021. The current version of the AI Act will next have to be adopted by the European Parliament (Parliament).

In January 2022, DLA Piper reported on an announcement of a new initiative, as part of the UK's National AI Strategy, to shape the way organisations and regulators develop technical standards for artificial intelligence ("AI"). The initiative, the AI Standards Hub ("Hub"), was highlighted as a collaborative effort between the Alan Turing Institute, the British Standards Institution, and the National Physical Laboratory, in partnership with the UK Government, to lead the way in developing standards that could be used across all sectors and jurisdictions. On 12 October, in their latest update, the Alan Turing Institute announced that the hard work of the collaborators was finally complete and that the Hub was ready for interaction. While still early in its use, the Hub already contains an array of resources that will allow its users to understand and help shape the role of standards in the development of AI and best practices. The primary goal of the Hub is to advance trustworthy and responsible AI through a focus on standards that can be used as part of governance and innovation tools and mechanisms.

A new kind of lawyer is coming to court -- one that's powered by artificial intelligence. Next month, a "robot" lawyer, which tells defendants what to say via bluetooth, plans to fight two speeding tickets in court, according to USA Today. This marks the first time AI will be used in court, Joshua Browder, CEO of DoNotPay, the startup behind the project, told the outlet. Although the company isn't making any of the details, including the identities of the defendants, public, they told USA Today that one person will argue their case in person while another will do so over Zoom. DoNotPay bills itself as "the home of the world's first robot lawyer," and says its mission is to "level the playing field and make legal information and self-help accessible to everyone," per its website.

If you use AI for any form of talent decision-making in your organization and it results in discrimination, whether it is by you or introduced by the AI, you are the one who is liable. When it comes to verifying the ethical nature of AI, this could be just the start of a global ripple effect. AI has the power to do a lot of good, but working on big data comes with risks. I've spent over 20 years as a workforce strategist scaling teams for some of the largest major projects in the world and have witnessed firsthand the impact of not having visibility of the skills and capabilities of my people. I've seen first-hand the amount of potential that was being wasted on our people and our business, which motivated me to develop an independently audited ethical talent AI.

The global economy is increasingly dependent on knowledge workers to meet the needs of public and private organizations. While there is no single definition of knowledge work, organizations and industry groups still attempt to measure individuals' capability to engage in it. The most comprehensive assessment of capability readiness for professional knowledge workers is the Uniform CPA Examination developed by the American Institute of Certified Public Accountants (AICPA). In this paper, we experimentally evaluate OpenAI's `text-davinci-003` and prior versions of GPT on both a sample Regulation (REG) exam and an assessment of over 200 multiple-choice questions based on the AICPA Blueprints for legal, financial, accounting, technology, and ethical tasks. First, we find that `text-davinci-003` achieves a correct rate of 14.4% on a sample REG exam section, significantly underperforming human capabilities on quantitative reasoning in zero-shot prompts. Second, `text-davinci-003` appears to be approaching human-level performance on the Remembering & Understanding and Application skill levels in the Exam absent calculation. For best prompt and parameters, the model answers 57.6% of questions correctly, significantly better than the 25% guessing rate, and its top two answers are correct 82.1% of the time, indicating strong non-entailment. Finally, we find that recent generations of GPT-3 demonstrate material improvements on this assessment, rising from 30% for `text-davinci-001` to 57% for `text-davinci-003`. These findings strongly suggest that large language models have the potential to transform the quality and efficiency of future knowledge work.

Although AI-based systems are increasingly being leveraged to provide value to organizations, individuals, and society, significant attendant risks have been identified. These risks have led to proposed regulations, litigation, and general societal concerns. As with any promising technology, organizations want to benefit from the positive capabilities of AI technology while reducing the risks. The best way to reduce risks is to implement comprehensive AI lifecycle governance where policies and procedures are described and enforced during the design, development, deployment, and monitoring of an AI system. While support for comprehensive governance is beginning to emerge, organizations often need to identify the risks of deploying an already-built model without knowledge of how it was constructed or access to its original developers. Such an assessment will quantitatively assess the risks of an existing model in a manner analogous to how a home inspector might assess the energy efficiency of an already-built home or a physician might assess overall patient health based on a battery of tests. This paper explores the concept of a quantitative AI Risk Assessment, exploring the opportunities, challenges, and potential impacts of such an approach, and discussing how it might improve AI regulations.