Emerging AI Tech: A Threat to Corporate Reputation?
Technology is a powerful tool for business.
Recent technologies have enabled businesses to communicate directly and even more authentically with customers through conversational marketing chatbots and social media. These methods allow for analysis of huge datasets to uncover new product and marketing insights. They streamline business processes and increase staff productivity.
While the revolutions in internet commerce, mobile connectivity, and social networking have all brought their own challenges—from data privacy concerns to the need to maintain cyber-security—the disruption they’ve caused has created incredible opportunities for savvy businesses.
However, a new range of artificial intelligence technologies are able to create high-quality, convincing fake videos, articles, and imagery pose a new and emerging threat to corporate reputation.
These deviant technologies are becoming increasingly sophisticated, easy to use, and accessible to non-experts.
3 AI Threats to Corporate Reputation
Three recent developments are of particular importance to those charged with protecting and improving corporate reputation:
- Deep Fake Videos
This downloadable software maps faces and turns them into lifelike videos, allowing the words, facial movements, and mannerisms of one individual to appear to be those of another. This Bloomberg demonstration shows the ease with which words can be put into the mouths of Presidents Obama and Putin, how John Oliver’s lines can become Stephen Colbert’s, and even how convincing mannerisms and facial tilts can be created.
Created by Uber software engineer Philip Wang, this website generates convincing artificially generated portraits. Every time the website is loaded, it creates a new—entirely fake—person. From formal corporate headshots to Instagram-style vacation selfies and school photos, the images are hyper-realistic. All the people below were created by the service and don’t exist. Keep hitting refresh to keep being freaked out!
- OpenAI's Text Generator
A Silicon Valley research institute backed by Elon Musk and Peter Thiel has created an artificially intelligent program capable of writing convincing articles with minimum human supervision. The Guardian calls its output “plausible newspaper prose,” while the BBC was so concerned it only quoted the text generator’s output as images, lest the text be scanned by search engines and treated as actual BBC reporting.
Much of the coverage of these technologies has focused, understandably, on their potential impact on politics and elections. But put to malicious use, they will pose a business reputation risk as well.
How AI Poses a Threat
Consider the following dark scenarios and the potential damage each poses:
- Thousands of critical reviews appearing on Amazon, TripAdvisor, or Trustpilot for a product or service—each review unique, well-written, and supported by lifelike imagery
- High-quality video emerging of a CEO making offensive and derogatory remarks about a minority group recorded at a private event
- A cache of classified, corporate branded documents leaking that reveal serious product flaws
These aren’t entirely new challenges. Fake reviews are a well-documented problem, and urban myths have always circulated about businesses and their leaders. But businesses have been able to adopt strategies to combat these. Fake reviews, for example, can be identified by their repetition or poor writing style, while reused images can be found through reverse image searches.
But advances in AI and related technologies change the potential quality and quantity of fabricated digital assets that can be created—and the ease and speed with which they can be produced. Fake written reviews will be harder to spot, fabricated videos will appear and sound more convincing than ever, and artificially generated documents will be harder to tell apart from what's genuine.
The potential sources of these threats are numerous: activists looking to discredit corporations whose business operations they object to, desperate companies willing to engage in dirty tricks to gain competitive advantages, and, like some hackers, those doing it simply for the "lulz."
Ways to Win on Trust
When a new threat to corporate reputation emerges, it is natural to try to downplay its potential impact. We might hope that our customers and stakeholders would recognize these as fakes, or that our businesses wouldn’t be targeted. But these technologies will pose a significant reputational risk.
They emerge at this critical time when business needs to win back trust. Globally, we found that just 38.5% of the public trusted business to do the right thing in 2018, so there is little general goodwill to fall back on. Attacks on reputation using these AI approaches will target existing reputational weaknesses—the CFO of an investment fund allegedly boasting of evading tax payments, or a technology CEO belittling the role of women in their organization.
As communication and reputation leaders, we need to begin to act proactively to prevent these new challenges. Imagine if an incriminating video emerges of a member of your c-suite:
- How long would it take to consider that the video might be an artificially generated fabrication?
- Where would you find a technical expert able to analyze the video and determine if it was, indeed, a fake?
- What explanation would you use to explain the attack to your customers, investors, employees, and regulators?
- Would they be inclined to believe you—or dismiss the explanation as an excuse for poor behavior by leadership?
There can be no "perfect" protection from a novel attack on reputation like this. But a strong reputation can help mitigate the risks these technologies pose and buy time to respond to them. Businesses with strong or excellent reputations enjoy greater trust, gain the benefit of the doubt at times of crisis, and can find ambassadors willing to speak positively about them.
Those businesses that enter a crisis with only an average or even weak reputation won’t enjoy these advantages and will find it much harder to respond.
Director of Consulting