Elisabeth Strenger’s Post

The next time you encounter someone handwaving away AI risks..."oh, AI isn't there yet..", "It is just a statistical word predictor...", "meh...AGI? it can't even do some arithmetic reliably yet..." direct them all here https://lnkd.in/gWRcKJGA These same, feeble, first version LLMs with plenty of warts are already causing many many harms...today, here and now... Want some examples? click away and discover for yourself... #aigovernance #aisafety #llms #airisk #governance

Did you know about the Ai Incident Database : https://lnkd.in/dBfUkx3C ? This database collects AI related incidents published in journals and makes them public for everyone. Next step for the incident data base : how to get companies to share incidents (not wait for the incident to be published in journals) ? Does anyone know of a company that uses the incidentdatabase or would be willing to disclose AI incidents? Remark : Thanks to our speakers Heather Frase, Ph.D., CAMS and Patrick Hall to respond positively to my request to present at the Berkeley Open Innovation Seminar and discuss "Why sharing AI incident is a must ?" Thank to the Berkeley Open Innovation Seminar (Henry Chesbrough,Agnieszka Radziwon and Mehdi Montakhabi) to support the organization of this debate on "Why sharing AI incident is a must ?" #OpenInnovation #Seminar #AI #Incident #CompetitiveAdvantage #BerkeleyInnovation

Keep track of AI harms and near misses in the real world with the AI Incident Database https://lnkd.in/eEbpfN2R #ai

I was searching for articles on AI to support my proposal and found an interesting database that captures events/incidents involving AI. Check it out at https://lnkd.in/gJJq9wGA It does trigger the necessity of having governance and standards to avoid more harmful AI released or shall we be reminded of the possibility of the movie The Matrix or Terminator becoming real? #ai #artificialintelligence #ethicalai #responsibleai

https://lnkd.in/egH37Qw If you are looking for TRA .. this is a good source of information of AI incidents that will stir your negative side about AI. #ai

More from the theme "how AI can fail (humanity) and what we've learned". The AI Incident Database is very interesting, and has incidents starting from 1996. For example, it has the famous 2016 incident when TayBot was introduced by Microsoft and failed within 24 hours when it became extremist due to social media data ingestion). The database contains more than 600 AI related incidents, with top 10 going to the expected companies (Tesla, OpenAI, Microsoft, Facebook, Google, Amazon). Have a look. #artificialintelligence #ai #aisafety

As I know the first, and if not, one of the first reliable sources of incidents impacting and/or involving the use of AI https://lnkd.in/dQTqcf8M. From my point of view it's an excellent source of information on the risks associated with AI.

AI is becoming crucial in global competition, especially in autonomous and decision-making systems for battlefield applications. However, current #AI models are vulnerable to cyberattacks and lack robustness for mission-critical decisions. A partnership between ORNL and Vanderbilt University aims to develop secure, dependable AI systems for #nationalsecurity applications. National Security Sciences at ORNL

The Wall Street Journal, new AI risk - "Data Poisoning," where bad actors spread faulty data that AI models incorporate when training. Enter RAG - Retrieval Augmented Generation (RAG). #rag couples up-to-date, actual enterprise data with LLMs trained on publicly available to ensure the most accurate results - regardless of any faulty model training inputs. Learn more at https://lnkd.in/epcV9m7a #genai #datasecurity #rag https://lnkd.in/ehzGmqdk

The IT landscape has seen the evolution of Artificial Intelligence (AI) over the years, and even more in the last 12 months. Can AI be leveraged for managing networks too? A whitepaper by Ramesh N G, Microland's Senior Director of Digital Network Security, examines how AI networking - utilizing AI techniques and algorithms - can help enterprises. https://lnkd.in/g87GPHCN #Microland #Ai #ArtificialIntelligence #NetworkSecurity #DigitalSecurity #AICloud #AIOps #AInetworking #IntelliginiNetOps