Opinions expressed by Entrepreneur contributors are their very own.
Key insights
- Companies can not view data as endlessly renewable. We are facing a “data liability gap” – the difference between the information you think that you’ll be able to access and the information you’ll be able to actually get better in a usable format.
- AI systems depend on complete historical data sets to learn and proper errors. Therefore, lost or corrupted data may result in erroneous or incorrect conclusions.
- Many executives assume that cloud availability equals data protection. In reality, cloud providers operate the service, but partners and customers are still chargeable for data backup and recovery.
In recent years, the company world has adopted the mantra that data is at all times renewable. Basically, people viewed storage as a utility and bandwidth as something that can at all times be there. Backup was viewed similarly to insurance. Since the arrival of artificial intelligence, all of this has been proven incorrect. As corporations increasingly depend on AI and predictive analytics, frightening possibilities are opening up.
We are currently facing a “data liability gap,” which is the difference between the information an organization believes it may access and the information it may actually get better in a usable format. Because AI systems rely heavily on old data to learn and proper their very own errors, everlasting data loss is not any longer just an operational threat; It is now such a serious matter that it could have to be mentioned within the annual reports. If it’s lost as a result of negligence, the responsible employees could also be fired as a result of the chance to the corporate’s popularity.
For generations, the C-suite viewed data protection as something much like data recovery. Their goal was to bring the systems back online as quickly as possible after the failure of key operational equipment. The concept of Recovery Time Objective (RTO) focused totally on speed. The principal goal was to get the servers up and running again.
AI has completely modified the sport. Instead of worrying about how long your systems have been online, AI systems care about historical data. An AI language model will face serious problems if it seems that records from the primary five years of the corporate’s existence have been destroyed or corrupted. This signifies that the prediction algorithms are missing essential historical data needed to make conclusions. In the worst case, misleading or completely incorrect conclusions are reached.
Unrecoverable data may end up in high costs
Many CFOs will agree that data is the essential raw material needed within the AI industry. Data integrity can also be essential and is a very important backbone to keep up operations. A producing company would suffer greatly if it discovered that a small amount of its raw materials had been destroyed from its warehouse. If this were to occur, there could be a serious investigation and an adjustment to the general value of the corporate.
2025 research by ExaGrid with Enterprise Strategy Group found that just one% of organizations are capable of get better all of their data after a ransomware attack.
However, when corporations discover that critical data they need starting in 2020 is broken beyond repair, the response could also be something like, “It’s a shame, but we have to move on.” This is despite the incontrovertible fact that the knowledge contained in the information would have enormous long-term value to the corporate.
The reasons for data loss should not only cyber attacks. It is estimated that in Microsoft 365 systemsIn 2025, roughly 30.2% of corporations lost data, a rise of 17.2% from 2024. This was due, for instance, to erroneous deletions or employees leaving who didn’t hand over the information properly.
Why “shared responsibility” will not be a very good attitude
The “availability myth” is a nasty strategy that’s unfortunately utilized by many managers today. When this happens, data is assumed to be secure just because the cloud during which it’s stored is secure easily available. Grant Crough, founder and CISO of LEAP Strategy, described this well as he said“Microsoft operates the service, but partners and customers remain responsible for data backup and recovery.”
Because corporations don’t understand the shared responsibility system well, they’ve suffered severe data loss. Modern Microsoft infrastructures are typically designed to guard corporations from hardware failures, not errors attributable to users. When ransomware attacks a system, it modifies every copy in a SharePoint library.
The only reliable protection against that is an independent backup that follows the 3-2-1 rule and consists of three copies (two media types and one external copy). Many executives mistakenly consider that Microsoft offers this, although it does not the case.
What the C-Suite must do in the long run
For an extended time, data management was focused on the server room or the IT team. Things need to vary and the boardroom must take more responsibility. The C-suite must concentrate on making data available indefinitely moderately than focusing their efforts totally on disaster recovery.
For example, executives have to concentrate on things like the proportion of their data that could be restored to a very good state and whether or not they have backups which can be resistant to strong attacks. If no answer could be given, it proves that there’s a serious weakness in the corporate. As the AI race continues to advance, the winners won’t be those with essentially the most data; They shall be those who’ve built indestructible protection systems for his or her data.
Key insights
- Companies can not view data as endlessly renewable. We are facing a “data liability gap” – the difference between the information you think that you’ll be able to access and the information you’ll be able to actually get better in a usable format.
- AI systems depend on complete historical data sets to learn and proper errors. Therefore, lost or corrupted data may result in erroneous or incorrect conclusions.
- Many executives assume that cloud availability equals data protection. In reality, cloud providers operate the service, but partners and customers are still chargeable for data backup and recovery.
In recent years, the company world has adopted the mantra that data is at all times renewable. Basically, people viewed storage as a utility and bandwidth as something that can at all times be there. Backup was viewed similarly to insurance. Since the arrival of artificial intelligence, all of this has been proven incorrect. As corporations increasingly depend on AI and predictive analytics, frightening possibilities are opening up.
We are currently facing a “data liability gap,” which is the difference between the information an organization believes it may access and the information it may actually get better in a usable format. Because AI systems rely heavily on old data to learn and proper their very own errors, everlasting data loss is not any longer just an operational threat; It is now such a serious matter that it could have to be mentioned within the annual reports. If it’s lost as a result of negligence, the responsible employees could be fired as a result of the chance to the corporate’s popularity.
For generations, the C-suite viewed data protection as something much like data recovery. Their goal was to bring the systems back online as quickly as possible after the failure of key operational equipment. The concept of Recovery Time Objective (RTO) focused totally on speed. The principal goal was to get the servers up and running again.
