Saidot

Read our full guide and get answers on how to ensure compliance with the AI Act, and also to these questions: ✔️ What is the AI Act, and who does it affect? ✔️ What does the Act require of organisations? ✔️ When will it come into force? ✔️ What penalties and enforcement will the Act introduce? ✔️ Which AI systems and use cases are classified as high-risk? ✔️ How can organisations comply with the Act? ✔️ How can organisations implement AI governance frameworks for it? Give it a read: https://lnkd.in/eb7EZj-V #EUAIAct #AIAct #AIcompliance #ResponsibleAI #AIsafety

✔️ We're one of the 100+ signatories of the AI Pact. 👇 Swipe left to see what we're committed to as a part of this Pact. #AIPact #AIAct #EUAIAct #AIcompliance

🙌 We're proud to be involved in AISym4MED - Horizon Europe research program. The target of the program is to enable medical research and development through a platform that supports safe collaboration between data providers, enabling synthetic data generation, model developers and researchers. AISym4MED research consortium presented the program at the European Big Data Value Forum 2024 in Budapest, Hungary with great interest and feedback from the event participants. Our Policy Manager Maria Koskinen held presentation at the General Assembly meeting, sharing insights on EU AI Act and relevant standards that set requirements to the platform design and development. #EBDVF2024 #MedicalResearch #FutureOfHealthCare #HealthTech #AI Iiris Lahti

🧤 #WhatsYour6thFinger: Could it be cyber security risks? AI can be exploited by attackers to gain control over computer systems, reveal sensitive information, or to cause malfunctioning. For example, hackers can do this through manipulating training datasets, social engineering, or exploiting computer vulnerabilities. 👉 Learn more about AI’s hidden risks: https://lnkd.in/eHZzavju #CyberSecurity #ResponsibleAI #AIsafety #AI

Read our blog and learn more the concept of AI hallucinations and how to manage them as part of a responsible AI governance strategy: https://lnkd.in/eGkqdfBQ #WhatsYour6thFinger #AIhallucination #Hallucination #AIrisks #GenerativeAI

🧤 #WhatsYour6thFinger: Could it be biased outcomes? AI bias refers to distorted AI outputs that reflect human bias originating in skewed training data, for example. AI algorithms often too have biases in information processing that can lead to ethical problems, for example in recruitment tools. 👉 Learn more about AI’s hidden risks: https://lnkd.in/e6hz48Vq #AIrisks #Hallucination #ResponsibleAI #GenerativeAI

Our AI Governance Intern Vesna asked ChatGPT (using the DALL·E model) to generate images of Berlin. Can you see what's wrong with these images? It generated some interesting versions of the city’s famous landmarks: Multiple TV towers and a hybrid version of the Reichstag Building and Berlin Cathedral. These hallucinations don’t have real-world consequences, as they're easy to recognise for those familiar with the city, but they allude to a deeper issue. Some of generative AI's mistakes can be challenging to spot, and in many real-world applications, hallucinations can have harmful effects. Read Vesna's blog post, in which she explores the concept of AI hallucinations and how to manage them as part of a responsible AI governance strategy. This quick guide includes: 🔮 What hallucinations are in the AI context, and why they occur. 🔮 Why hallucinations are a problem and the vital ethical considerations. 🔮 How to manage and mitigate the risk of hallucinations. Give it a read: https://lnkd.in/eaCrHkAC (P.S. See content credentials for the images in Vesna's original post through the link in the comments.) #WhatsYour6thFinger #Hallucination #AIhallucination #AIrisks #GenerativeAI

🧤 #WhatsYour6thFinger: Could it be harmful and toxic content? AI systems may generate harmful or inappropriate content and spread hate speech and misinformation. For example, individuals' voices and appearances can be recreated through deepfake technology to manipulate others to believe something that is not true. 👉 Learn more about AI’s hidden risks: https://lnkd.in/ezYFuT_Z #ResponsibleAI #AIrisks #Deepfake #GenerativeAI

We're happy to share #WhatsYour6thFinger has been featured again in Spanish, this time in El Confidential and FayerWayer. 🇪🇸 Gracias, six times over, for writing the story and amplifying the message about AI's risks and the importance of identifying and mitigating them. 🔗 Read the full stories through the links in the comments. #AIrisks #ResponsibleAI #AIsafety United Imaginations FI

🧤 #WhatsYour6thFinger: Could it be privacy infringements? AI systems can pose privacy risks by processing and analysing personal data without consent and sufficient safeguards, potentially violating individuals' privacy. Personal information can be inadvertently included in the AI-generated output and generative AI models can identify individuals by combining multiple data sources. 👉 Learn more about AI’s hidden risks: https://lnkd.in/eXKXUBcS #AIrisks #Privacy #ResponsibleAI #GenerativeAI