AI’s hunger for more training data has led companies like Adobe and Zoom to attempt changes in their Terms and Conditions, seeking access to their users’ cloud data. The resulting backlash has prompted these companies to reconsider their actions. However, this trend is not isolated. Alphabet, for instance, has initiated product features that scan users’ Google Drives without prior notice or an opt-out option, raising concerns about data privacy and security.
Some may downplay this recent incident, arguing that Google has always had questionable privacy policies. However, the surge in data collection attempts could spell the end of the comprehensive online services we’ve grown accustomed to.
The Convenient Cloud
While some cloud services, such as webmail, have been around since the mid-1990s, the true boom in cloud services began with the launch of Dropbox in 2007. Dropbox’s user-friendly file synchronization service marked a significant shift as cloud services replaced everyday activities like copying files between computers. This evolution has integrated ecosystems like iCloud, revolutionized business processes like Quickbooks Online, and democratized technology access through platforms like the Google Play Store.
These activities transitioned from our local systems and corporate servers to cloud servers housed in massive data centers. Software as a Service became the norm for companies. Your organization could concentrate on revenue-generating tasks, while all the software for support services, like HR, marketing, and Finance, could be handled by those who could operate them much more efficiently.
Free and Free
Even better, many of the consumer versions came at no apparent costs. Advertisement, data analytics, and the pressure for businesses to adopt familiar systems were enough to justify the zero price. Crutial, till the advance of AI, data analytics did not involve the content of the customer’s data, e-mails, and video conferences. Thus, the loss of privacy in metadata was an acceptable price for consumers.
Companies are more averse to having their data analyzed or their employees viewing ads. Thus, they were willing to pay for services.
We should accept that there is no such thing as a free lunch. We have to pay for privacy if we want privacy.
Business Risk Cloud Service
Till now, the data for training usage has only hit non-paying consumer users. Yet, the changes to Adobe’s and Zoom’s Terms and Conditions should be a warning. They only backed off from the goal of using customer data as training data after a considerable outcry from artists and professionals, respectively. Thus, it is only a question of time before a company feels secure enough in its market leadership not to back off.
The Google Drive issue from above has come without warning, and it ignored the setting to keep it out. Consequently, the IT department, management, and boards must be aware of any change in terms and conditions. After all, any of them might mean that the company could start to index or process your files shortly.
The issue also highlights the importance of fighting against shadow IT and for compliance. Any authorized service should have vetted terms and conditions. The same won’t be valid for unauthorized cloud services. Thus, keeping your internal data out of the training sets of AI services suddenly depends on enforcing compliance with the restrictions on your shadow IT.
The End of The Cloud?
Cloud computing relies on the implicit promise that the service provider can supply a better solution than an on-premise IT department ever could. However, with the increasing attacks on our privacy and data security, it is time to reconsider whether the monetary savings are worth the increased risks. In particular, it might be time to rethink public cloud and software as a Service strategy for risk-averse companies and regulated organizations.
The rise of AI might change the way we interact with software. After all, we don’t rouge AIs to leak our personal and business documents.