Data Security Fears: Congress Bans Staff Use of Microsoft’s AI Copilot

[Caute_cautim]

Hi All

Wow, that was quick!!

U.S. House Bans Staff from Using Microsoft’s Copilot Amid Data Security Concerns – The AI coding assistant, recently released to the public, is deemed a risk due to potential data leakage – Learn more about the controversy and the future of AI in government workplaces.

https://www.hackread.com/us-congress-bans-staff-microsoft-ai-copilot/

Regards

Caute_Cautim

[APigman]

I have been reading about Microsoft's use of AI. They are hosting their own engine and the reality is Microsoft owns most org's data anyway. Their transparency around how it is used at GitHub was very well done compared to every other company. They seem to be taking a more responsible approach than most companies I've seen. Data is not being sent to ChatGPT, though they are using ChatGPT's technology. Could the data used to train it result in disclosure? Maybe - but Microsoft is used regulatory agencies around the world - they have to be thinking about how to keep those folks from getting skittish.

[Caute_cautim]

@APigman   Unfortunately they say one thing and do another, Microsoft needs to demonstrate good ethics and governance practices throughout their AI models.   With some of the things you can do to prompt Copilot to do, and recently cause erroneous output or believe it is the Messiah is not good.  Now that the Chinese have been successful in attacking Microsoft systems, they need to change the culture significantly - they have been put in position of trust with a vast amount of Government agencies around the world, a lot of whom are driven by big discounts and best value for money. 

There needs to be significant development in protecting AI models from various attacks including data poisoning, disruption and the ability to influence output too.   They have an uphill journey to undertake, as we know once trust is eroded it can be a significant effort to regain that lost ground, if not there will be many reservations.

Regards

Caute_Cautim

[APigman]

There is definitely a bell curve to grading how companies use AI. Microsoft
has been more transparent than most. There are a lot of companies who are
flat out reckless. There definitely need to be more controls in general and
it makes me wonder if the data poisoning attacks will move faster than the
protections given the broad attack surface. The genie is out of the bottle
- we can only hope to grow with it.

[Caute_cautim]

@APigman     Yes, I agree, I posted this piece on under governance, but it appears even indigenous people have issues with the Microsoft hack too even here in New Zealand, it has widespread issues.

https://www.linkedin.com/pulse/ai-media-weekly-commentary-m%2525C4%252581ori-indigenous-april-10-dr-...

Regards

Caute_Cautim