By Michael S. Borella –

Once upon a time, patent eligibility was not controversial or difficult to understand. Then along came Alice Corp. v. CLS Bank, and with it the Supreme Court’s bright idea to replace statutory clarity with metaphysical hand-waving about so-called abstract ideas. The result has been a decade of improvisation, where examiners and judges try to divine the essence of a claimed invention by sniffing for abstraction like sommeliers. One of the most unintuitive outcomes is that an independent claim can be patent-eligible while a dependent claim that narrows it with more detail can be found ineligible.

To make the issue concrete, let’s consider two claims about routing data packets in a network.

1.          A computer-implemented method for dynamically routing data packets in a packet-switched network, the method comprising:

receiving, by a network interface of a router, a data packet comprising a header and a payload;

parsing, by a packet-processing module executed by a processor of the router, the header to extract a destination network address and a priority value;

analyzing, by the packet-processing module, a packet-type indicator contained in the payload to identify a packet type;

selecting, from among a plurality of output ports, an output port based on the destination network address, the priority value, and the identified packet type, in combination with current link-utilization metrics of the output ports stored in a routing table maintained in memory;

updating the routing table to record packet-transmission statistics associated with the data packet; and

transmitting the data packet via the output port.

This claim is about as technical as it gets. It involves a processor, a routing table, memory, and measurable performance metrics. The improvement is concrete; namely, dynamic routing based on link utilization and packet type to improve throughput. Under U.S. case law and patent office procedure, this claim should be a textbook example of non-abstractness. It is the kind of claim that can sneak through the Alice gauntlet.

Now, let’s add a dependent claim.

2.         The computer-implemented method of claim 1, wherein the packet type corresponds to a banking transaction, the payload comprising encrypted banking transaction data formatted according to a predefined financial messaging protocol, and wherein selecting the output port further comprises prioritizing the data packet for transmission over a low-utilization output port.

Here, the claimed invention still processes packets and it still selects output ports based on a combination of factors. The only change is that the payload now contains encrypted banking transaction data, and the routing algorithm prioritizes the packets for transmission on a low-utilization output port. Not only does this balance load but it allows these types of packets to be processed faster due to the relationship between network link utilization and latency.

And yet, under Alice, that dependent claim is probably ineligible. Why? Because the payload includes banking transaction data. Suddenly, one can argue that the invention is directed to the abstract idea of intermediating financial transactions. Never mind that the machine is still transmitting the same bits through network interfaces. The difference in the semantic content of the bits is determinative.

To be fair, the patentee still has an opportunity to argue that claim 2 in its full context (which includes all limitations of claim 1) recites an inventive concept that overcomes its allegedly abstract nature. However, the Alice test provides a number of pathways for contending that claim 2 fails to provide such an inventive concept. For example, courts may conclude that the type of narrowing of claim 2 is a field-of-use restriction rather than an inventive concept, or that speeding up financial transactions addresses a fundamental business need rather than a technical improvement. They may also view the limitations of claim 1 as well-understood, routine, and conventional, whereas this avenue would not be available when claim 1 is considered in isolation.

These two claims illustrate how the Alice framework can turn a concrete, machine-based process into an abstract idea merely by referencing the data to which the process is applied. Under this logic, there are cases in which the more specifically you describe a computer-implemented method, the greater the risk of ineligibility. A broad, generic claim about “routing data packets” may be safe. But narrow it down to “routing banking packets with latency reduction,” and it magically transforms into an abstract business concept. This inversion of logic punishes claim language precision, the very thing patent law is supposed to reward.

It is difficult to overstate how irrational these results can be. Routing packets? Not abstract. Routing packets containing medical data? Abstract. Encrypting chat messages? Not abstract. Encrypting stock trades? Abstract. In these cases, the technology has not changed. Instead, the message encoded in the bits determines whether the machine performing the task is concrete or abstract. It is like deciding that a hammer is a physical tool when driving nails to construct a table but an abstract idea when used to make a treasure chest.

This doctrinal absurdity doesn’t just distort legal reasoning. It also warps applicant behavior. Under Alice, applicants have a perverse incentive to claim their inventions more broadly than they otherwise would. Normally, a careful drafter writes claims to highlight concrete technical features such as precise hardware interactions, data structures, and control logic. But thanks to Alice, there are situations where the more you explain about what your system actually does, the easier it is for an examiner or judge to find in it an abstract idea. So applicants may hedge their bets by de-emphasizing the actual invention they wish to protect and keeping claims focused on the broader technical features. In some cases, this involves omitting any discussion of the actual intended use of the technology from the patent application completely. In other words, Alice can reward exactly the kind of overbroad claiming that its advocates continue to complain about.