As interest in and implementation of ChatGPT and generative artificial intelligence more broadly ramps up, so too does concern about privacy and
trust and calls for regulation.
Earlier this month, Italy became the first Western country
to block ChatGPT over unease about privacy controls. The country’s data
protection agency has given OpenAI, maker of the generative AI chatbot, a list
of demands to meet by April 30. On Tuesday the company announced
new privacy controls that seem to be in direct response to some of Italy’s
concerns, including the ability for users to disable chat history, thereby allowing
them to choose which conversations can be used to train ChatGPT’s models.
In the United States there has been talk, but no action,
about the need for regulation, although Monday
CNBC reported a tech advocacy group representing Microsoft, Adobe, IBM and
Oracle is advocating for rules governing the use of AI in national privacy legislation.
Meanwhile, the launch of new tools and services using
ChatGPT continues apace. And travel is certainly one of the verticals that is
seeing a flurry of activity. In March, Expedia
and Kayak became two of the first brands to release plug-ins for ChatGPT and
less than two weeks later Expedia
updated its mobile app to include ChatGPT for trip planning. Trip.com
has also connected OpenAI’s API to its mobile app, and dozens of other
companies are rolling out integrations seemingly daily.
Some industry experts are sounding the alarm about the
convergence of these two issues – legitimate concerns about data privacy and
trust and a powerful technology that is becoming ubiquitous.
“With the scale of power that’s now available to us, the
scale of catastrophe is also raised,” said Trevor Butterworth, co-founder and vice president for communications and governance for Indicio,
a developer
of open-source decentralized identity.
And the risk is most evident in the types of applications that have captured the attention of the travel industry, where the buzz is about using generative AI for personalization and fast, frictionless experiences for shopping and booking. With the launch of plug-ins in March, ChatGPT can now access external data sources, creating more opportunities but also more risks for privacy and security.
“Because you can have an intelligent agent - whether it’s a
travel agent, a financial agent – that is capable of accessing your data,
learning from it, understanding in a sense … in terms of correlations and
patterns and behavior,” Butterworth said.
“And if that is owned by somebody else, and that is hacked
or you’re hacked, your security is compromised. That’s a whole other level of
dystopia.”
Solving ChatGPT’s trust issue
In a position
paper recently published by Indicio, Butterworth and co-author Karl
Schweppe, head of innovation at digital product studio Bay Tree Ventures, propose
a solution: verifiable credentials – defined by the W3C as “a
tamper-evident credential that has authorship that can be cryptographically
verified” - used with a decentralized identifier communications protocol
(DIDComm). Used in tandem, the authors propose, these technologies enable secure,
trustworthy person-to-AI interaction – mitigating risk and maximizing a
seamless experience for both companies and consumers.
Subscribe to our newsletter below
In
travel specifically, they write, the use of verifiable credentials would mean
consumers can use ChatGPT agents – what they call a Chat Valet – to handle tasks
such as booking flights and hotels according to preferences and then automatically
paying for those bookings through secure, digital communication with the
traveler’s bank.
“We’ve spent 20 years gaining a sense of trust in the internet
and in the brands and experiences we use on the internet, with about 80%
confidence and 20% ‘this is tolerable,’” said Schweppe.
“With the new models of chat, AI and those interfaces, we
have to rebuild a whole new infrastructure of trust. … In financing we have payment
rails. We will need identity rails and trust rails for chat. But it’s going to
be really challenging because it’s moving so fast.”
Butterworth and Schweppe have both been working with verifiable
credentials for some time, but they say it was the release of OpenAI’s plug-ins and
API in March that made them realize ChatGPT creates the perfect use case for this
decentralized identity technology.
“It’s the only solution that can really manage the need for
continuous verification between each party in the interaction,” Butterworth
said.
“I recognize my ChatGPT agent as being mine, and it
recognizes me as the owner. Whoever we are dealing with – the hotel for example
- recognizes that the ChatGPT agent is my agent and that it’s not interacting
with someone else. And when we exchange information, this process of
verification is continuous. It applies zero trust relentlessly in order to
ensure our data is secure.”
Schweppe predicts chat could eventually replace a website as
the consumer’s interface with a brand, eliminating traditional logins and
creating the need for a new model of identity verification.
“Up until this point we’ve always had walled gardens
where you log into a website, and they protect your identity,” he said.
“With chat, they aren’t there in the same way anymore. The chat
is going to be your interface that can then communicate and integrate with
multiple services that are not necessarily walled in the same way. So the identity piece becomes so much more important.”
Mitigating the risk
of “fake everything”
Jamie Smith, an expert on self-sovereign identity who writes
about the topic on his Customer
Futures blog, agrees that digital trust is critical right now because AI is
creating “an explosion of fake everything.”
“In a world where I can’t tell what is real and fake, we’ll
need digital wallets to help us prove who we are and for us to make sense of
what we are seeing,” Smith said.
We will need identity rails and trust rails for chat. But it’s going to be really challenging because it’s moving so fast
Karl Schweppe - Bay Tree Ventures
“Yesterday it was fake reviews on Tripadvisor, tomorrow it
will be you’ve taken my money for a package trip, but I get off the plane and there
is no tour guide, there is no bus to collect you, there is no hotel. All the reviews
are there but along with the website, the photos, the phone call you had with the
agent – none of it is real. AI is surfacing incredible opportunity, but we have
a fake everything problem.”
But the need for consumers to trust the brands and content
they interact with in chat is just one piece of the puzzle. In order for the
full power of AI to be activated to provide consumers with customized, frictionless
shopping and booking experiences, they will need to give it their personal preferences
and data – and trust that it is secure. Conversely, brands will need to know the
consumers they are interacting with in chat are “real” and to recognize that verifiable
credentials create an opportunity to – finally – build a fully-connected travel
experience.
“Now more than ever the travel ecosystem is looking for more
collaboration, more connected journeys, more value and that’s the super power
of verifiable credentials – portability,” Smith said.
“And in a world where personal AI is going to get closer to
us and become increasingly more valuable, it needs to be trained on data it can
trust, and individuals need to know what’s happening to that data. And verifiable
credentials create a great supply chain of who produced what data, when and
how.”
Schweppe said companies should realize “this is early-adopter territory so brands that make an impact now do have an opportunity to
get ahead of others that are slow to move. … For that reason, it’s imperative to
stay aware of what’s going on and … begin to think what strategy will work.”
And with heightened concerns globally about data privacy and
scrutiny of “big tech” such as Facebook, Google, Amazon and Apple, Butterworth
said recognition of the risks associated with emerging AI solutions is critical.
“Across different industries and within verticals we are
seeing companies take data privacy and security seriously. And [with verifiable
credentials] there’s not a trade-off – it comes with the benefit of verifiable
data,” he said.
“So the win-win aspect of decentralized identity and verifiable
credentials is really apparent to a lot of companies. And in the marketplace,
it will be who do you trust with your data? Can you control your data? That’s
going to be a very simple, powerful message that I think will animate a large
section of the marketplace both from business and from consumers.”
Phocuswright Europe 2023
Generative AI such as ChatGPT is set to revolutionize travel. Leaders from
Trip.com, Microsoft and Kayak tell us how the technology will change everything from the back of the house to the front.