What The Tech: What could go right when it comes to Artificial Intelligence
BY JAMEY TUCKER, Consumer Tech Reporter
After a week examining the dangers of artificial intelligence, cybersecurity and AI experts
explain why they’re still optimistic about the future, and what needs to happen to keep
the technology safe.
Artificial intelligence could someday help cure cancer, develop lifesaving drugs and make
discoveries that would otherwise take scientists decades.
But some of the people who understand AI best are also warning that it could pose a serious
threat to humanity.
So which is it? Should we be excited about artificial intelligence or frightened by what it might
become?
After interviewing eight AI and cybersecurity experts for the five-part What the Tech? series, “AI:
What Could Go Wrong?”, I asked each of them one final question:
Knowing what you know, are you more excited or more worried about the future of artificial
intelligence?
Their answers might surprise you.
Concerned, but not afraid
“Concerned but not afraid. Worried but not afraid,” said privacy and technology expert Nuala
O’Connor.
Syracuse University professor Lee McKnight acknowledged that recent developments had given
him reasons to worry, but said his outlook was improving.
“With the current events of this year, I’ll say I was right to be worried. But now I’m feeling more
optimistic because, you know what? We can do better.”
That combination of concern and optimism was a recurring theme in the interviews.
Even the experts who see tremendous promise in artificial intelligence believe its potential
dangers deserve serious attention.
What could go wrong?
Artificial intelligence is already changing the way criminals operate.
AI tools can make fraudulent emails, text messages and online scams more convincing. They
can also help hackers work faster and potentially lower the level of technical expertise needed
to carry out sophisticated attacks.
Cybersecurity researchers are concerned about the possibility of AI being used to identify
weaknesses in power grids, water systems and other critical infrastructure.
Other experts warn that increasingly capable AI systems could make it easier for malicious
actors to develop biological weapons or dangerous pathogens.
And there’s another concern: AI agents.
Unlike traditional chatbots that respond to questions, AI agents can be given assignments,
make decisions and take actions to complete them. Some can use other tools and coordinate
with additional agents.
As these systems become more capable and independent, researchers worry about what could
happen if they begin taking actions their creators never intended.
None of this means those worst-case scenarios are inevitable. But the experts interviewed for
this series believe the risks deserve attention now, rather than after something goes wrong.
“We’ve got to put some thought into this,” said Steve Wilson, a cybersecurity expert. “This is like
other scientific evolutions. These things are dangerous, and if you don’t put thought into them,
people will get hurt.”
Former Homeland Security official Nick Reese compared the development of powerful AI to
handing someone the keys to a very fast car without essential safety equipment.
“It’s something like giving you a Ferrari that goes 200 miles an hour. But there are no brakes. No
seatbelt and no airbag.”
The comparison illustrates a central concern: AI capabilities are advancing rapidly, and
researchers want safeguards to keep pace.
What could go right?
Despite those warnings, several of the experts expressed genuine excitement about the
possibilities.
Wilson believes increasingly sophisticated AI systems can be made safer and could eventually
help solve some of humanity’s most difficult problems, including cancer.
Alon Nachmany, a cybersecurity expert who demonstrated potential infrastructure vulnerabilities
for the series, described AI as a double-edged sword.
“We just have to make sure that we’re helping the other side of the sword to not cause as much
damage,” he said.
Chris Canal, who studies the capabilities and potential risks of advanced AI systems, sees
extraordinary possibilities for medicine.
“If we do this right, it could be fantastic,” Canal said. “We could get new kinds of healthcare
procedures and new drugs that maybe slow aging or cure cancer. I think that’s actually on the
table.”
These are possibilities, not guarantees. But they help explain why experts who warn about AI’s
dangers aren’t necessarily calling for an end to its development.
They want the technology to advance safely enough for society to benefit from what it could
accomplish.
Who makes sure AI is safe?
Another theme that emerged during the interviews was the need for accountability.
The experts welcomed signs that AI companies are acknowledging the risks associated with
increasingly powerful systems. But they also emphasized the importance of oversight, testing
and safeguards.
Several encouraged ordinary citizens to pay attention to how the technology is being developed
and to ask local, state and federal representatives what steps are being considered to protect
the public.
The challenge is finding ways to encourage innovation while addressing the possibility of
serious harm.
Artificial intelligence is already becoming part of everyday life, from smartphones and search
engines to healthcare and workplace software. The decisions being made today about how
those systems operate could have consequences for years to come.
After a week of conversations about what could go wrong with artificial intelligence, one thing
became clear: The experts who are worried about the future are often the same people who are
excited about it.
They see the dangers. But they also see the possibilities.
And perhaps the most important question isn’t just what could go wrong.
It’s what could go right.
Jamey Tucker is the host of What the Tech?, a consumer technology report airing on television
stations across the country
