Thread by @MiekeEoyang

Now, from #HewlettVerify a conversation with @Microsoft's Brad Smith & @chinmayiarun on "Safety and Security In a Digital World." Will tweet...

@Microsoft @chinmayiarun Smith starts off talking about Christchurch attack.

"We've created a world where any human being can broadcast live to the planet."

And we saw this abused in a horrible way.

@Microsoft @chinmayiarun Social media thinks about responsibility in circumstances like this.

We need to do things differently. It takes a while to figure out what to do.

Are there criteria before someone engages in live broadcasting?

Can think about collaboration, crisis response differently.

Rules around the world need to change. Watch for announcement from Paris May 15, where PM Ardern will be.

Now asking about ethical principles for AI:

Facial recognition and AI for the military are the first two policy issues to emerge.

Facial recognition tech yielding benefits, but can be misused or abused.

Notes that facial recognition tech has challenges with people of color, women and can result in discrimination.

Facial recognition tech improves with those who have the most data. But agglomeration of data can result in a race to the bottom.

Need a floor and need a law on what facial recognition policies should look like.

Certain technologies (facial recognition) that you might not want to provide to authoritarian governments, but there might be entities within that nation that could obtain technology with certain controls.

But must think through the nuances.

Development of a Digital-Geneva Convention?

Civilians should be protected in cyberspace. But what form of protection? Maybe a new treaty someday. But it should bring together a number of norms and rules.

Need to give all norms more support and more momentum.

Paris Call has a wide range of stakeholders as signatories. It's an example of what's needed to promote cybersecurity.

It has:

1) a multilateral approach to diplomacy

2) it's a multi stakeholder document

3) Also need public diplomacy. (but it's hard to get people interested in Tallin 2.0, but easier to understand cyber-Geneva.)

Smith making an analogy to arms control and chemical weapons treaties.

Need to have a broader conversation with the public.

But should we approach this via treaty, given US reluctance to sign treaties?

Paris Call also lacking US & India.

"I have no doubt that the US will sign the Paris Call. And I have no idea when that will happen."

Paris Call is a norm that needs the support of the USG and is consistent with US values.

Even though multilateralism not in vogue in DC, remains hopeful that it will happen.

What are you doing on elections?

We launched a defending democracy initiative. We have to protect campaigns, candidates, think tanks. And have to defend against misinformation.

Have been taking legal action to disrupt foreign activity against campaigns.

Also working on protection of voting. stay tuned for public announcements of same.

Ideas: creating a more trusted voices for news. Ex. supporting @NewsGuardRating. Needs more collaboration from tech.

@NewsGuardRating What should the law and governments do to protect against disinformation?

Particular interest in @MarkWarner's white paper on social media regulation: qz.com/1345888/mark-w…

@NewsGuardRating @MarkWarner @granick elicits a discussion about limitations of livestreaming.

@NewsGuardRating @MarkWarner @granick I asked about what government needs to do to get better at prosecuting cyber crimes, based on @comey's remarks yesterday.

@BradSmi's answer: Need a multidisciplinary approach and must hire data scientists and have the in the room as part of investigation.

@NewsGuardRating @MarkWarner @granick @Comey @BradSmi @BradSmi also talks about the @Microsoft threat intelligence center and developing a strategic approach to dealing with cybercrime, not just whack-a-mole. Important to disrupt the bad actors not just prosecute low level folks.

@NewsGuardRating @MarkWarner @granick @Comey @BradSmi @Microsoft Coming back to @granick's question, there was a good back and forth about whitelisting journalism outlets, and the risk to limitations that doesn't allow victims and witnesses to upload the crimes, rather than the perpetrators.

@NewsGuardRating @MarkWarner @granick @Comey @BradSmi @Microsoft @SpauldingSez asks about protecting rule of law and preventing people from undermining legitimacy of institutions.

1) protect people who are under attack and also deter the bad actors.
2) Then also how do we get to trustworthy sources of news?

Also must reform the revenue model for journalism, given loss of classified ad dollars.