Apéro Digital Infographic III - Challenges of Generative AI for Business

Addressing the Challenges GenAI-appropriate risk management strategies need to be adopted into regular processes. Upskilling strategies covering both technical & non-technical aspects should be developed, and addressed to all hierarchy levels. Change management strategies can support AI governance implementation plans more effectively. What is GenAI AI systems that produce text, images, video or other types of data They learn patterns from training data & produce new data based on prompt input Examples of GenAI systems include ChatGPT, Dall-E, Co-Pilot & Llama Assessing reliability & transparency of GenAI is particularly challenging GenAI Use Cases Virtual Assistants & Chatbots Text summarization / translation / generation Semantic search in (large) documents Generating programming code Text, Image, video generation & editing for entertainment / art (Semi)automatic content creation for marketing Regulatory Challenges Lack of clear guidelines pushes organizations to “wait and see”. Fears of copyright infringement widely exist, even though legal risks may be lower than perceived. Legal demands for explainability are high yet hard to guarantee. Need for constant refinement to stay compliant with current standards. Apéro Digital Challenges of Generative AI (GenAI) for Business Technical Challenges Accountability methods for GenAI, such as privacy-preserving design, are underdeveloped. Confidentiality is a source of concern, due to the risk of leaking confidential information. Efforts needed to build responsible GenAI systems, such as applying ethics assessment frameworks, are underestimated. Concerns that implementing trustworthiness safeguards may decrease the GenAI performance are present. Organizational Challenges Companies feel pressure to find use cases for GenAI, rendering the trustworthiness of the system an afterthought. Organizations & customers expect GenAI to “make life easier”, yet are often unwilling to pay for necessary controls. Negative consequences on the work environment are expected, as some professions may risk to “die out”. Domain experts’ experiences are challenged, making them often the least acquainted with AI technology. APÉRO DIGITAL III – OCTOBER 2024